Firecrawler

**Judgment:** High-utility, well-scoped web scraping/crawling skill built on the Firecrawl API, suitable as a primary tool for fetching live web content, documentation, and structured page data. It is powerful but depends on an external API key, credit quotas, and responsible use of web scraping. **What it does well** - Fetches pages as **clean markdown**, including JS-rendered content, which is ideal for downstream summarization or analysis. - Takes **full-page screenshots** for visual inspection or archival. - Performs **structured data extraction** from pages via a JSON schema and optional prompt, enabling product info extraction, listings, etc. - Provides **web search + content retrieval**, offering a path to current information (subject to Firecrawl’s plan/limits). - Can **crawl and map documentation sites**, making it strong for framework/library onboarding and building local knowledge bases. **Key risks & limitations** - **External dependency:** Relies on Firecrawl’s service availability, performance, and API stability. Any outage or quota limit directly impacts this skill. - **Cost/credit constraints:** Free tier is 500 credits; each page/screenshot/search/crawl page costs 1 credit. Large crawls can exhaust quotas quickly if limits are not set conservatively. - **Legal and policy compliance:** Web scraping can conflict with site terms of service, robots.txt, or regional regulations. The skill does not enforce compliance; users must ensure lawful and ethical usage. - **Content fidelity & structure:** While it handles JS content and schema-based extraction, the quality of extracted data depends on site structure and Firecrawl’s parsing/LLM behavior; complex or irregular pages may yield incomplete or noisy data. - **Not suited for authenticated/private data:** It’s designed for public web content; using it on login-protected or sensitive endpoints is risky and may violate terms or privacy expectations. **Recommended scenarios** - When you need **current public web information** from specific URLs or search queries (e.g., “scrape this blog post and summarize it”). - When you need **consistent markdown representations** of sites or docs to feed into other OpenClaw skills (analysis, Q&A, embedding, etc.). - When you want to **crawl technical documentation** (frameworks, APIs, SaaS tools) so the system can later answer in-depth technical questions. - When you must **extract specific fields** (titles, prices, features, metadata) from product or content pages using a JSON schema. - When you require **visual snapshots** of pages (UX review, compliance evidence, simple before/after comparisons). Use this skill when you explicitly need **live web access, structured scraping, or doc crawling**, and you’re able to manage API keys, credit usage, and compliance with target sites’ policies.